Analysis of the Spatial and Temporal Locality in Data Accesses

Abstract

Abstract. Cache optimization becomes increasingly important for achieving high computing performance, especially on current and future chip-multiprocessor (CMP) systems, which usually show a rather higher cache miss ratio than uniprocessors. For such optimization, information about the access locality is needed in order to help the user in the tasks of data allocation, data transformation, and code transformation which are often used to enhance the utilization of cached data towards a better cache hit rate.In this paper we demonstrate an analysis tool capable of detecting the spatial and temporal relationship between memory accesses and providing information, such as access pattern and access stride, which is required for applying some optimization techniques like address grouping, software prefetching, and code transformation. Based on the memory access trace generated by a code instrumentor, the analysis tool uses appropriate algorithms to detect repeated address sequences and the constant distance between accesses to the different elements of a data structure. This allows the users to pack data with spatial locality in the same cache block so that needed data can be loaded into the cache at the same time. In addition, the analysis tool computes the push back distance which shows how a cache miss can be avoided by reusing the data before replacement. This helps to reduce cache misses increasing therefore the temporal reusability of the working set.

Guided Prefetching Based on Runtime Access Patterns

“…For acquiring access patterns we developed an analysis tool [14] which currently detects both access chain and access stride. The former is a group of accesses that repeatedly occur together but target on different memory locations.…”

An Interactive Graphical Environment for Code Optimization

“…The base of the analyzer [11] is a memory reference trace which records all runtime memory accesses. By applying algorithms, often used in Bioinformatics for pattern recognition, it finds the affinity and regularity between the references, like access chain and access stride.…”

“…This includes a cache visualizer [9] which demonstrates cache problems and the reason for them, a data profiler which collects information about global cache events, a cache simulator [10] for delivering runtime cache activities, and a pattern analysis tool [11] for discovering the affinity among memory operations.…”

A spectral clustering approach for multivariate geostatistical data

“…Methods to account for spatial correlation in a predictive modelling task (classification or regression) have been proposed, for instance, in references [5,16,30,43,44,50]. Approaches accounting for spatial correlation in a descriptive problem have proposed, for example, in references [29,45].…”